Currently, computer data (for example, data stored on disk volumes) are backed up either on pre-set schedules, or in the event of a known impending disaster (for example, a loss of power, an imminent flood, etc.). This leaves any new or modified data fully vulnerable between backups, because it is not possible to recover data more recent than the last executed backup.
On the other hand, each executed backup uses both computing cycles and storage resources. Therefore, although backing up frequently is safer in terms of data integrity, it can become very resource expensive. Traditionally, a compromise has been reached for each protected system by balancing the time between backups (the window of vulnerability) and the resources needed or available to perform or hold the backup data.
In contemporary computing environments, information from external sources concerning the stability of the system is often available. For example, a malicious code threat assessment system can provide information regarding a current threat level. However, today's backup systems do not leverage external states to enhance data backup management.
What is needed are methods, systems and computer readable media to improve backup management, by automatically adjusting the frequency and/or data granularity of backups responsive to states external to the backup system.